Modal
Free tierHigh-performance AI infrastructure with sub-second cold starts and instant autoscaling
Key strengths
Facts verified against modal.com on
Recent changes
- Starting price: $30/mo → $0/mo
- Plans updated: Starter plan is $0 + compute/month (includes $30/month free credits, 3 seats, 100 containers, 10 GPU concurrency). Team plan is $250 + compute/month (includes $100/month free credits, unlimited seats, 5000 containers, 50 GPU concurrency). Enterprise is custom pricing. Compute billed per second: e.g. H100 SXM5 at $0.001097/sec, CPU at $0.0000131/core/sec, Memory at $0.00000222/GiB/sec.
Get alerts when Modal changes
Free. We'll email you when pricing, the free tier, API access or open-source status changes.
Modal — Technical Overview
Modal is a serverless cloud compute platform purpose-built for AI/ML workloads, offering GPU and CPU infrastructure billed per second with no idle costs. Founded in 2022, Modal provides a Python-native SDK and API that lets developers define, deploy, and scale containerized functions directly from code — no Kubernetes, no Dockerfiles, no infrastructure YAML required.
Key technical characteristics:
- Per-second billing on all compute: e.g. H100 SXM5 at
$0.001097/sec, CPU at$0.0000131/core/sec, memory at$0.00000222/GiB/sec - Instant autoscaling to thousands of GPUs and millions of sandboxes, driven by a custom container runtime and fast scheduling engine
- Multi-cloud infrastructure spanning 20+ clouds, with availability on AWS Marketplace and GCP Marketplace
- Workload support across inference serving, model training, batch jobs, and code sandboxes
- Observability integrations with Datadog and OpenTelemetry
- Identity & access via Okta integration
- Security posture: SOC 2 Type II compliant and HIPAA ready
- Open source client/SDK with 1,282 GitHub stars
- API available for programmatic control of deployments and infrastructure
Starter plan includes $30/month in free compute credits, 100 containers, and 10 GPU concurrency slots. Team plan ($250 + compute/month) scales to 5,000 containers and 50 GPU concurrency. Enterprise pricing is custom.